Autopilot AI in NetSuite is here. The hard part is earning it.
Updated: Aug 28
Ollie Holt and I spent yesterday at #SuiteConnect in #London at the #InterContinental London - The O2, listening to the keynotes, watching the demos, and (like everyone else) having the same recurring thought: ERP conferences now contain more AI than coffee, or tea (I'm a tea man), which is saying something.
But here’s what surprised me.
Under the noise, this was not really an “AI is coming” event. It was a “your ERP is becoming the control layer for AI” event. That is a very different proposition, and it is why the most memorable line of the day was not a feature name. It was a framing:
Evan Goldberg talked about NetSuite’s ambition to be an “autopilot” for businesses, not another copilot you chat to on the side.
If you work in finance, IT, or transformation, you will know why that landed. Copilots answer questions. Autopilot changes outcomes, ideally with you still in charge of what matters.
So, what did SuiteConnect London actually tell us about where NetSuite is heading, and what should UK customers pay attention to now?
The UK backdrop: AI spend is happening, whether we like the hype or not
One useful reality check: AI is not a side bet for leadership teams anymore. In February, EY reported that 96% of UK CEOs said they will be investing in AI technology in the next 12 months.
That statistic is not a licence to buy tools impulsively. But it explains the mood: CFOs are being asked to fund AI, govern AI, and still get the month end close done on time. CIOs are being asked to make AI safe, while also making it fast. COOs are being asked to operationalise it without breaking controls.
SuiteConnect, at its best, tried to connect that pressure to practical product direction.
What felt real: finance automation moving from “help” to “execution”
The single most finance relevant announcement stream sits in and around the close.
NetSuite 2026.1 introduces NetSuite Intelligent Close Manager, described as a new dashboard portlet that uses transaction data and AI driven exception detection to give a central view of tasks, KPIs, transaction amounts, and exceptions across areas like AR, AP, and accounting. It is designed to help teams identify issues earlier and keep close activities on track.
This is not just another checklist with nicer colours. The intent is a shift towards continuous oversight: surfacing what is not matching expectations while there is still time to fix it, rather than discovering problems on day three of close.
Two related details matter for real world finance teams:
The portlet is positioned as something you enable, with tasks then created automatically based on transaction activity and enabled features.
Exception Management enhancements in 2026.1 include payment risk detection, flagging changes to critical vendor data fields near payment events to help identify potential fraud faster.
Why this matters to CFOs and controllers: it is a move from retrospective variance explanation to earlier interruption of risk and error. That reduces days lost to detective work, and it increases confidence that the close is accurate because you had fewer surprises.
Now for the realism.
If your close is currently held together by spreadsheets, informal heroics, and institutional memory, AI will not magically turn it into a controlled process. Intelligent Close Manager will shine a very bright light on gaps in task ownership, inconsistent transaction discipline, and poorly governed master data. That is painful, but it is also useful.
In other words, everyone wants autopilot until they remember someone still has to tidy the data.
AI in NetSuite is not just the close: matching, narratives, and metrics
The broader AI feature batch that NetSuite has been announcing across its SuiteConnect tour is a “many small wins” strategy, provided you implement them with care.
NetSuite has highlighted AI powered bank transaction matching to increase auto match rates and reduce manual review, plus AI generated report narratives to turn dense business data into action, and AI assisted pricing summaries that consolidate inventory, costs, and sales data into a single narrative.
For subscription and SaaS finance teams, NetSuite also called out advanced subscription metrics, including committed monthly recurring revenue metrics and cohort analyses, framed as executive ready reporting and early churn indicators.
None of these replace a finance function. They reduce friction. The organisations that win will be the ones that treat friction reduction as a programme, not a scattered set of toggles.
Governed AI and MCP: the most strategically important but least “wow” topic
The most important announcement in London, in my view, was not a single feature. It was the growing clarity around how NetSuite intends to connect internal ERP data with external AI models safely.
NetSuite has adopted the Model Context Protocol (MCP) as part of its AI Connector Service, positioning it as a standards based way for supported AI clients to access and interact with NetSuite data and functionality.
At SuiteConnect London, NetSuite announced additions to that connector story, including:
AI Connector Service Companion, positioned as a finance grade experience, with a prompt library of more than 100 finance focused templates aligned to NetSuite terminology and structures, plus reusable “skills”, and preconfigured role patterns mapped to roles like CFO, controller, AR analyst, AP analyst, and treasury analyst.
MCP Apps, aimed at bringing structured NetSuite style interfaces (report picker, record picker, prompt library) into AI assistants so users are not relying purely on free text prompting. These are planned for release as part of a future MCP Standard Tools SuiteApp release, available through the SuiteApp Marketplace.
Expanded AI connector support for NetSuite Analytics Warehouse, extending AI reach beyond transactional ERP data into historical, analytical, and third party data stored in the warehouse.
For a transformation leader or CIO, the significance is simple: the future is not one model. It is many models, many agents, and many use cases, and the enterprise problem is controlling them.
This is where governed integration beats AI theatre.
If you can connect a model to ERP data with role based permissions, audit trails, and repeatable prompts, you can move from experimentation to something you can defend to an auditor, a board, and an unpleasant surprise from your internal controls team.
And yes, there were demos that leaned into external AI assistants. Coverage of the event described use cases like pulling overdue accounts data, generating dashboards, and even using multimodal input to log photographed inventory items into NetSuite.
The point is not the specific demo. The point is the architectural direction: AI as an execution layer that you can govern.
As a NetSuite partner, we’re treating AI connections like any other integration: valuable only if the controls are watertight. NetSuite’s AI Connector Service is built on the Model Context Protocol (MCP) and uses OAuth 2.0 with NetSuite roles and permissions, so administrators can explicitly authorise which roles can access the MCP server and installed tools, and adjust that scope over time. NetSuite also states its service encrypts data in transit and at rest, and it operates regional data centres (including Europe) on OCI—so data residency becomes a shared responsibility the moment prompts or outputs touch an external AI service. Our baseline: least‑privilege AI roles, separate read vs write tools, approvals for any record updates, and routine reviews of logs/audit trails.
Integration is being repositioned as a business capability, not just a technical cost
A quieter but commercially meaningful piece of news is the NetSuite Integration Platform.
NetSuite announced the availability of a low code, AI powered integration platform designed to automate and manage cross application workflows and connect NetSuite to third party systems using natural language and prebuilt AI integrations. It also explicitly positions governance in the integration layer, with role based access, audit trails, monitoring, and MCP server security.
For UK customers, the availability note matters: NetSuite stated the Integration Platform is now available to customers in the UK and Ireland.
From a solution provider partner perspective, this is where delivery gets interesting. Most ERP pain is not inside the ERP. It is “ERP plus everything else”:
Payroll systems that do not agree with finance
Inventory systems that do not agree with operations
Customer data that exists in four places, none of them trusted
If integration moves towards low code design with AI assistance, it can shrink delivery time. But the same realism applies: low code does not remove the need for ownership, data definitions, and monitoring. It just makes the consequences of poor design arrive a bit faster.
If, like me, you try your hand at anything, this part is for you.
From a partner delivery standpoint, SuiteCloud Developer Assistant is one of the more practical AI additions: it runs in VS Code via the Cline extension and lets us scaffold new SuiteCloud projects from a prompt, producing boilerplate plus unit test files and setup guidance. It supports SuiteScript 2.1 work and SDF‑centric builds, including creating SDF XML custom objects and updating manifests, and it can generate unit tests with mocked dependencies. That combination usually means faster first drafts, better baseline test coverage, and fewer avoidable SDF packaging mistakes, which reduces rework when we hand work into version control and release pipelines. Caveat: Oracle is explicit that outputs need human review and should be tested outside production; never deploy assistant‑generated code straight into a live account.
SuiteCloud Developer Assistant is available now as part of the SuiteCloud development tooling ecosystem
It is accessed via Visual Studio Code, using the relevant NetSuite extensions and configuration (not directly inside the NetSuite UI)
It supports SuiteScript 2.1, SDF projects, and developer workflows today
NetSuite Next (pure dead brilliant)
On NetSuite Next, the next real evolution of NetSuite and ERP in general. The point made was not just a prettier UI, although it looks like the much needed update to bring back the wow factor; it is a new way of getting work done. Oracle describes Ask Oracle as a natural-language assistant that lets users search, navigate, analyse and act across the entire NetSuite dataset, returning context-aware answers, visualisations and explained reasoning. NetSuite’s own Next pages make the operating model explicit: users can kick off agent-led workflows inside Ask Oracle, track progress, and still keep agents inside the same roles, permissions and approval flows that already govern the account. On timing, Oracle’s launch announcement is clear that Next will be available initially for North American customers within the next 12 months. UK rollout timing remains something to validate, not assume, but there should be previews available in 6 months with production available within 12. And how do you switch to NetSuite Next? Mini implementation...data migration...NO! A simple click of a button and you're in the NetSuite Next edition seamlessly. More on NetSuite Next to come, but lets just say we're all excited here and you should be too!
Customer proof points: growth stories still matter more than vendor slides
The best moments of SuiteConnect events are usually the customer stories, because they force the conversation back to operations.
One of the London proof points was Octopus Energy, with commentary in the run up to the event highlighting how the business is using NetSuite to help manage growth. If you have followed Octopus, you will know their operating model is as much technology platform as energy supplier, with Kraken referenced publicly as part of the group. Kraken Technologies
That matters because it grounds the “unified data” message in reality: scale is not just volume, it is complexity across entities, geographies, products, and controls.
What I would tell UK CFOs, CIOs, COOs, and transformation leaders this morning
If you are a NetSuite customer, or evaluating NetSuite, the practical takeaway from SuiteConnect London is not “turn on AI”. It is this:
Start with the close, because it is where discipline shows up. Intelligent Close Manager and exception management will reward teams that already have defined ownership, clean vendor controls, and reconciliations treated as ongoing hygiene rather than a month end event.
Treat AI governance as a design requirement, not a phase two. The connector story, MCP, role mapped access patterns, and a prompt library designed for finance is a signal that NetSuite expects customers to connect external AI. Your job is to make that safe and repeatable.
Simplify before you automate. AI will not rescue messy processes dressed up in better interfaces. It will just produce answers and actions faster, which is only helpful if your underlying data and workflow logic is sound. The organisations that win will be the ones with decent data, decent process discipline, and enough courage to remove complexity rather than rebrand it.
Watch NetSuite Next, but do not confuse roadmap with readiness. NetSuite is clear about the direction: Ask Oracle at the centre, agentic workflows, and a switch to the Next experience without disrupting customisations. Initial availability statements have been North America first, with wider rollout following, so UK customers should stay curious and ask for specifics rather than assuming immediate access.
The biggest cheer of the day was not for a model, a protocol, or an abstract vision. It was for the idea that finance might spend less time chasing and more time deciding. If NetSuite can deliver that without weakening controls, it will earn the autopilot metaphor.
The question is whether we, as customers and partners, are willing to do the unglamorous work that makes autopilot safe: data ownership, workflow discipline, and change management that actually sticks.
What is the one part of your finance or operations process you would not trust “autopilot” with yet, and why?
Contact John directly to find out more.
